Nanomaterials for Nanoscience and Nanotechnology part 6 pdf

Under the impact of an incident electron, the electrons bounded to the atoms may be excited either to a free electron state or to a unoccupied energy level with a higher energy.. Second,

terized by the quantized energy levels and the associated electronic states Under the impact of an incident electron, the electrons bounded to the atoms may be excited either to a free electron state or to a unoccupied energy level with a higher energy The quantum transitions associated with these excitations will emit photons (or x-rays) and electrons such as secondary electrons, Auger electrons and ionized electrons, these inelastic scattering signals are the finger prints of the elements that can provide quantitative chemical and electronic structural information

Figure 3-22 shows the main inelastic processes that may be excited in high-energy electron scattering [6, 43] When a fast electron passes through a thin metal foil, the most noticeable energy-loss is to plasmon oscillations in the sea of conduction elec-trons For an ideal case in which the electrons can move ªfreelyº in the sea, the system can be treated as an electron gas This case is best represented by aluminum, in which the outer-shell electrons can be considered as free electrons The negatively charged particles are mixed with nuclei, forming a solid state plasmon ªgasº The resonance frequency of this plasmon is directly related to the density of electrons in the solid The decay of plasmons results in the emission of ultraviolet light The cathodulumines-cence (C.L.) technique is based on the detection of the visible light, which is emitted when an electron in a higher-energy state (usually at an impurity) fills a hole in a lower state that has been created by the fast electron

Figure 3-22 Schematic one-electron energy level diagram plotted against the positions of atoms show-ing the characteristics excitations by an incident electron in a semiconductor material Here, E f is the Fermi level, E b the binding energy, and C.B., V.B and V.L are the conduction-band minimum, valance-band maximum, and vacuum level, respectively
E 1 a K shell excitation;
E 2 a single-electron excita-tion; C.L cathodoluminescence photon; P a plasmon.

Second, atomic inner-shell ionization is excited by the energy transfer of the inci-dent electrons, resulting in an ejected electron from a deep-core state (Figure 3-23a) Since only one inner-shell electron is involved in the excitation, this process is also called single-electron excitation The excitation introduces an energy-loss in the range

of a few tens to thousands eV which is the finger print of the element Since the inten-sity and threshold energy of the inner shell electron transition are determined not only by the binding energy of the atom but also by the density of states in the valence band, the energy-loss near edge structure usually carries some solid state effects, and this is the basis of analyzing the valence states of an element Analogous to C.L., the holes created at deep core states tend to be filled by the core-shell electrons from higher energy levels, the energy is released in the form of photons (or x-rays) (Figure 3-23b) The energies of the emitted x-rays are quantized and they are fingerprints of the corresponding elements and are used for chemical microanalysis The holes,

creat-ed by the ionization process, in deep-core states may alternatively be fillcreat-ed by the electrons from outer shells, the energy released in this process may ionize another outer shell electron (Figure 3-23c), resulting in the emission of Auger electrons Accompanying to these processes, second electrons can be emitted from the valence band The difference between Figures 3-23b and 3-23c is the emission of photon in the former and the Auger electron in the latter, and the two processes are complimentary

Figure 3-23 (a) Ionization of an atom bounded inner-shell electron by an incident electron, resulting in (b) X-ray emission and (c) Auger electron emission (see text).

In other words, the elements with higher Auger electron yields have lower x-ray emis-sion, and vice versa Thus, the Auger electron spectroscopy is more sensitive to light elements while EDS is to heavier elements

The last characteristic inelastic excitation is phonon scattering (or thermal diffuse scattering, TDS) [44], which is the result of atomic vibrations in crystals This process does not introduce significant energy-loss but produces large momentum transfers, which can scatter the incident electron into a high angular range This is a localized inelastic scattering process As will be shown in Chapter 4 the collection of TDS elec-trons can produce compositional sensitive images

3.6.1 Valence excitation spectroscopy

In studying of nanoparticles, it is necessary to probe the electronic structure of a single nanocrystal This is possible only with the use of a probe that is smaller than the size of the nanocrystal The valence excitation spectrum of a nanoparticle is most sen-sitive to its electronic structure and size effects [45, 46] The spectra can be acquired

in TEM and STEM using a fine electron probe The quantification of the spectra relies

on theoretical calculation The valence band excitation of a nanoparticle is most easily and accurately described using the dielectric response theory The impact of an inci-dent electron is equivalent to a time-depeninci-dent pulse, which causes transitions of valence electrons In classical dielectric response theory, an incident electron is treated

as a particle following a pre-defined trajectory, which is assumed not being affected by the interaction between the incident electron and the dielectric media, as shown in Figure 3-24 Electron energy-loss is a continuous process, in which the electron is decelerated due to the attractive force Fz= (± e)Ezowing to the field of the induced charges, resulting in energy-loss For a general case in which the incident electron is moving along the z-axis and under non-relativistic approximation, if the instantaneous position of the electron is denoted by r0= (x0, 0, z¢=vt), where x0is called the impact parameter, the energy-loss spectrum of an incidence electron due to surface plasmon excitation of a finite dielectric medium is calculated by [45, 47]

dP…o†

do ˆphve2

R1

1 dz0 R1

1 dz oIm{± exp [io(z¢±z)/v] Vi(r,r0)}

|r=(x 0 ,0,z), r 0 =(x 0 ,0,z¢) (3-12) where Vi(r,r0) is the potential due to the induced charge when a ªstationaryº elec-tron is located at r0= (x0, 0, z¢), i.e., it is the homogeneous component of V satisfying

Ñ2V(r,r0) = ±e…o†ee

for the dielectric media considered It is important to note that Vi(r,r0) is o-depen-dent The potential distribution in space is a quasi-electrostatic potential for each point along the trajectory of the incident electron The integral over z¢ is to sum over the contributions made by all of the points along the trajectory of the incident elec-tron Therefore, the calculation of valence-loss spectra is actually to find the solution

of the electrostatic potential for a stationary electron located at r0 in the dielectric media system We now use a spherical particle as an example to illustrate the applica-tion of this theory

Consider a point electron moving at a constant velocity v in vacuum along a trajec-tory specified by r0 = (x0,0,z¢), as schematically shown in Figure 3-24 For simplicity one assumes that the electron does not penetrate the particle In terms of spherical coordinates (r,,), r = (x0+ z2)1/2, cos  = z/r, r0= (x0+ z¢2)1/2and cos 0= z¢/r0 The potential due to the incident electron in free-space for r0> r is

Ve(r,r0) = ±4pe e

0jr r0j = ±4pee0r0 P

1 Lˆ0

PL mˆ0NLm(r/r0)LPLm(cos)

where PLmis the associated Legendre function and

NLm=…2 d0m†…L m†!

where d0m is unity if m = 0 and is zero otherwise If r > r0, then r and r0 are exchanged in Eq (3.149) The solution of Eq (3-13) consists of the field by the free point charge and the induced charge The potential outside the sphere has a form of

Vout(r,r0) = Ve(r,r0) ± e4pe

0

P1 Lˆ0

PL mˆ0ALm1

r (a/r)L+1

PLm(cos) PLm(cos0) cos(m) (3-15) and inside the sphere the potential is

Vin(r,r0) = ± e4pe

0a P

1 Lˆ0

PL mˆ0BLm(r/a)LPLm(cos) PLm(cos0) cos(m) (3-16) Matching boundary conditions:

Vout(r,r0)|r=a = Vin(r,r0)|r=a, and@Vout…r;r0†

@r |r=a = e(o)@Vin…r;r0†

@r |r=a (3-17) the ALmand BLmcoefficients are determined to be:

Figure 3-24 Excitation of a spherical particle by an exter-nal incident electron with an impact parameter of x 0 (x 0 > a) The radius of the sphere is a and its dielectric function

is e(o).

ALm= NLm L…1 e†

Le‡L‡1 (a/r0)L+1, and BLm= NLm 2L‡1

Le‡L‡1 (a/r0)L+1 (3-18) Substituting Eq (3-15) into (3-14), performing the analytical integral with the use

of an identity

R1

1 dt r0…L‡1†PLm…cos0†exp…iot† ˆ

2iL mjo=vjLKm…jox0=vj†

L m

(3-19) the spectrum is given by

dP

do = e

2

a

p2e0hv2

P1 Lˆ0

PL mˆ0MLm[Km(ox0/v)]2(oa/v)2LIm[ L…e…o† 1†L…e…o†‡L‡1] (3-20a) where

MLm= …2 d0 m†

From the energy-loss function, the resonance free frequencies of the surface plas-mons are determined by [50]

The homogeneous medium theory has been extended recently for the cases of an-isotropic dielectric medium, such as carbon onion structure and carbon nanotubes [48, 49] Shown in Figure 3-25a are a group of calculated EELS spectra for a carbon onion

Figure 3-25 (a) Calculated EELS spectra of a carbon sphere (radius = 10 nm) with graphitic onion-like structure as a function of the electron impact parameter x 0 (b) Experimentally observed EELS spectra from a carbon sphere (Courtesy of T Stöckli).

of radius 10 nm as a function of the electron impact parameter The calculation has to consider the anisotropic dielectric properties of the graphitic onion-like structure At small impact parameters, the excitation is dominated by volume plasmon, and at

larg-er impact parametlarg-ers, the surface excitation becomes dominant When the electron probe is outside of the sphere, the entire spectrum is the surface excitation This type

of calculation gives quantitative agreement with experimental observations (Figure 3-25b) and it can be used to quantify the dielectric properties of a single nanostruc-ture

3.6.2 Quantitative nanoanalysis

Energy dispersive x-ray spectroscopy (EDS) and electron energy-loss spectroscopy (EELS) in TEM have been demonstrated as powerful techniques for performing microanalysis and studying the electronic structure of materials [43] Atomic inner shell excitations are often seen in EELS spectra due to a process in which an atom-bounded electron is excited from an inner shell state to a valence state accompanied

by incident electron energy loss and momentum transfer This is a localized inelastic scattering process, which occurs only when the incident electrons are propagating in the crystal Figure 3-26 shows an EELS spectrum acquired from YBa2Cu4O8 Since the inner-shell energy levels are the unique features of the atom, the intensities of the ionization edges can be used effectively to analyze the chemistry of the specimen After subtracting the background, an integration is made to the ionization edge for

an energy window of width
accounted from the threshold energy Thus, the intensity oscillation at the near edge region is flattered, and the integrated intensity is domi-nated by the properties of single atoms This type of information is most useful in material analysis and the integrated intensity is given by

Figure 3-26 An EELS spectrum acquired from YBa 2 Cu 4 O 8 showing the application of EELS for quan-titative chemical microanalysis, where the smooth lines are the theoretically simulated background to

be subtracted from the ionization regions.

where I0(
) is the integrated intensity of the low-loss region including the zero-loss peak for an energy window
; A(
,) is the energy and angular integrated ionization cross-section In imaging mode, is mainly determined by the size of the objective aperture or the upper cut-off angle depending on which is smaller In diffraction mode, the angle is determined not only by the size of the EELS entrance aperture and the camera length but also by the beam convergence In general, the width of the energy window is required to be more than 50 eV to ensure the validity of Eq (3-21), and
= 100 eV is an optimum choice If the ionization edges of two elements are observed in the same spectrum, the chemical composition of the specimen is

nA

This is the most powerful application of EELS because the spatial resolution is almost entirely determined by the size of the electron probe The key quantity in this analysis is the ionization cross-section For elements with atomic numbers smaller than 14, the K edge ionization cross-section can be calculated using the SIGMAK pro-gram [43], in which the atomic wave function is approximated by a single-electron hydrogen-like model The ionization cross-section for elements with 13 < Z < 28 can

be calculated using the SIGMAL program For a general case, the ionization cross-section may need to be measured experimentally using a standard specimen with known chemical composition

3.6.3 Near edge fine structure and bonding in transition metal oxides

The energy-loss near edge structure (ELNES) is sensitive to the crystal structure This is a unique characteristics of EELS and in some cases it can serve as a ªfinger-printº to identify a compound A typical example is the intensity variation in the p* and * peaks observed in the C-K edge, as shown in Figure 3-27 Diamond is almost completely dominated by the * bonding, while the p bonding appears in graphite and

Figure 3-27 EELS C-K edge spectra acquired from diamond, amorphous carbon and graphite, respec-tively, showing the sensitivity of EELS to bonding in carbon related materials.

amorphous carbon The disappearance of the p* peak in C-K edge can be uniquely used to identify the presence of diamond bonding in a carbon compound It must be pointed out that the spectrum for graphite shown here was acquired when the incident beam parallel to the c-axis If a small size aperture is used, only a small portion of electrons corresponding to the p* peak is collected

In EELS, the L ionization edges of transition-metal and rare-earth elements usually display sharp peaks at the near edge region (Figure 3-28), which are known as white lines For transition metals with unoccupied 3d states, the transition of an electron from 2p state to 3d levels leads to the formation of white lines The L3and L2lines are the transitions from 2p3/2to 3d3/23d5/2and from 2p1/2to 3d3/2, respectively, and their intensities are related to the unoccupied states in the 3d bands Numerous EELS experiments have shown that a change in valence state of cations introduces a dra-matic change in the ratio of the white lines, leading to the possibility of identifying the occupation number of 3d orbital using EELS

EELS analysis of valence state is carried out in reference to the spectra acquired from standard specimens with known cation valence states Since the intensity ratio of

L3/L2 is sensitive to the valence state of the corresponding element, if a series of EELS spectra are acquired from several standard specimens with known valence states, an empirical plot of these data serves as the reference for determining the valence state of the element present in a new compound [51±59] The L3/L2ratios for

a few standard Co compounds are plotted in Figure 3-29a EELS spectra of Co-L2,3

ionization edges were acquired from CoSi2(with Co4+), Co3O4(with Co2.67+), CoCO3

(with Co2+) and CoSO4(with Co2+) Figure 3-29b shows a plot of the experimentally measured intensity ratios of white lines L3/L2for Mn The curves clearly show that the ratio of L3/L2is very sensitive to the valence state of Co and Mn This is the basis of our experimental approach for measuring the valence states of Co or Mn in a new material

Determination the crystal structure of nanoparticles is a challenge particularly when the particles are smaller than 5 nm The intensity maxim observed in the x-ray

or electron diffraction patterns of such small particles are broadened due to the crystal shape factor, greatly reduced the accuracy of structure refinement The quality of the

Figure 3-28 EELS spectrum acquired from MnO 2 showing the Mn-L 3 and Mn-L 3 white lines The five windows pasted in the Mn-L edge are to be used for extracting the image formed by the ratio of white lines.

high-resolution TEM images of the particles is degraded because of the strong effect from the substrate This difficulty arises in our recent study of CoO nanocrystals whose shape is dominated by tetrahedral of sizes smaller than 5 nm [60] Electron dif-fraction indicates the crystal has a fcc-type cubic structure To confirm the synthesized nanocrystals are CoO, EELS is used to measure the valence state of Co Figure 3-30

Figure 3-29 Plots of the intensity ratios of

L 3 /L 2 calculated from the spectra acquired from (a) Co compounds and (b) Mn com-pounds as a function of the cation valence.

A nominal fit of the experimental data is shown by a solid curve.

Figure 3-30 A comparison of EELS spectra

of Co-L 2,3 ionization edges acquired from

Co 3 O 4 and CoO standard specimens and the synthesized nanocrystals, proving that the valence state of Co is 2+ in the nanocrystals The full width at half maximum of the white lines for the Co 3 O 4 and CoO standards is wider than that for the nanocrystals, possibly due to size effect.

shows a comparison of the spectra acquired from Co3O4and CoO standard specimens and the synthesized nanocrystals The relative intensity of the Co-L2to Co-L3for the nanocrystals is almost identical to that for CoO standard, while the Co-L2 line of

Co3O4is significantly higher, indicating that the Co valence in the nanocrystals is 2+, confirmed the CoO structure of the nanocrystals

Figure 3-31 A schematic diagram showing energy-filtered electron imaging in TEM The conventional TEM image is recorded by integrating the electrons with different energy losses The energy-selected electron images corresponding to different characteristic energy-loss features are shown, which can be used to extract useful structural and chemical information of the specimen.

or electron diffraction patterns of such small particles